Monaco Report Reviewed

By John D. Perry, PhD

January 22, 2000

Executive Summary

The American Urological Association relied heavily on the report and recommendations from the First International Consultation on Incontinence (Monaco, 1998) in formulating their "Broad Guidelines for the Management of Urinary Incontinence," submitted to HCFA's MCAC (July, 1999).

Monaco used unusually strict exclusionary criteria when evaluating biofeedback research (but not for bladder retraining), and they excluded from consideration apriori most of the incontinence research funded by the U.S. government and conducted by the National Institute on Aging; research which is highly regarded in most professional circles.

Monaco relied heavily on a "systematic review" of research by Berghmans et al (1998) that rejected biofeedback as a viable therapy. But Weatherall (1999) has shown by meta-analysis that Berghmans should have drawn an opposite and favorable conclusion from the evidence he reviewed.

We present an in-depth analysis of the research in, and the conclusions of Berghmans' study; a new summary of important clinical trials; and of RCT research published after Monaco. We argue that the evidence does not warrant Monaco's rejection of biofeedback research, and therefore, their exclusion of biofeedback therapy from their incontinence treatment algorithms was inappropriate, and not in the best interests of Medicare.

Monaco Report Reviewed

By John D. Perry, PhD

The written testimony submitted to HCFA's MCAC by the American Urological Association[1] is based heavily on what has become known as the "Monaco Report". The AUA proposes several management algorithms in which biofeedback plays a significant role only in the treatment of the "frail elderly", generally considered poor candidates for surgery anyway. AUA claims that:

"These algorithms are based as far as possible on the evidence from the scientific literature and the opinion of the involved experts, including the experts of all relevant committees of the First International Consultation on Incontinence in Monaco, June 28-July 1, 1998. An effort has been made to make them easy to use and to reflect existing clinical practice. (p. 943)"

Indeed, the AUA committee has included many pages of the original as an Appendix (9b) to its testimony. Given the importance AUA has put on the Monaco Report, a closer examination of its methods and conclusions is in order.

The 1st International Consultation on Incontinence (Monaco, 1998)[2] concluded that there was NO evidence to support the use of Biofeedback [p. 591], NOR of electrical stimulation [p. 595], NOR even of physical therapies in general [p. 599][3]!

This conclusion -- reached by a world-wide committee of "academic" researchers -- flies in the face of the daily experience of thousands of Continence Clinics in the USA and elsewhere. It is at first difficult to understand how these researchers arrived at such a conclusion, until we examine their methodology in detail. We will focus especially on the biofeedback section.

TWO DIFFERENT INCLUSION CRITERIA

"Committee 14, Conservative Management [of Incontinence] in Women", was faced with the challenge of evaluating "any therapy that does not involve medical or surgical intervention." (p. 581). Since that includes a large and wide variety of research reports, they decided to restrict their search:

In view of the complexity of the task, the evidence reviewed is restricted to randomized clinical trials (RCTs[4]) in an attempt to differentiate specific treatment effects from a host of other variables. A summary of the … inclusion/exclusion criteria is given in Appendix 1. (p. 586) (Emphasis added)

The major consequence of this self-imposed limitation was to exclude from consideration virtually all of the research funded by the U.S. Public Health Service and the National Institute on Aging, and nearly two decades of work by the Laboratory of Behavioral Sciences under the direction of Dr. Bernard Engel.

The Committee intentionally excluded all prospective clinical series,[5] including the highly regarded NIA publications of Engel, Whitehead and Burgio, "in view of the complexity of the task". But was the task really that complex? They were left with 5 RCTs, and then eliminated one of them (Henderson and Taylor) on the grounds that their sample size (12) was too small. Having only 4 surviving RCTs lead to the conclusion that there was "limited evidence" available about biofeedback.

But a more relaxed Inclusion/Exclusion Criteria was employed in evaluating the literature on "Bladder Training (BT)"[6]. In that section they included both "randomized controlled and non-randomized controlled trials.

The Committee found that having only "five"[7] RCTs was an insufficient basis for evaluation; and therefore, "Because of the lack of RCTs, clinical series were also reviewed" (p. 613).

It is not explained why a pool of only five BT RCTs requires inclusion of additional clinical series research, but a pool of four BF RCTs is not entitled to the same consideration. This is more than a "half-empty vs half-full" debate, however, since the committee then decreed that "There is … evidence that bladder retaining is an effective treatment", while asserting that the evidence was too "limited" to call biofeedback effective. It was the committee that limited the evidence, and foreordained its own conclusion.

There is, of course, an obvious explanation for the lack of uniform standards that were used to compare biofeedback and other "physical therapies" against "bladder retraining", and that is that different members of the larger committee were responsible for the different sections. That is clearly suggested by the fact that different typefaces are used for the different tables.

The major impact of the International Continence Society's exclusion of "clinical series" when evaluating "biofeedback" was to prejudicially exclude from consideration the vast majority of the research funded by the U.S. Government, including many projects financed by the National Institute on Aging. Let us turn to examining the evidence that the Committee did include.

MONACO'S BIOFEEDBACK REVIEW

Only five studies involving any sort of biofeedback met the Committee's criteria of prospective RCTs. These included the following well-known reports:

1. Berghmans et al, 1996 (Netherlands)

2. Burns et al, 1993 (USA)

3. Castleden et al, 1984 (UK)

4. Glavind et al, 1996 (Denmark)

5. Taylor and Henderson, 1986 (USA)

Interestingly, one study that is commonly misclassified as being a "biofeedback" study (Ferguson et al, 1990) was correctly identified in Monaco as having used only "resistive devices", and it is discussed in that section.

On the other hand, another study, (Shepherd & Montgomery, 1983) has been misclassified by Monaco as a "resistive device" study, when in fact they did genuine Kegel-type biofeedback every day at home! (See below.)

Why did the Monaco Report include just the five studies it did? Perhaps it is only a coincidence, but these five studies are the same five described in a 1998 "systematic review" of biofeedback by Berghmans, Hendriks, Bo, Hay-Smith, van Waalwijk, and van Doorn (BJU, 1998). Two of Berghmans' co-authors in that project, Hay-Smith and Kari Bo, were also members of Committee 14, which wrote this report.

EXAMINATION OF THE RCT BIOFEEDBACK STUDIES

1. Berghmans et al, 1996

Berghmans et al used an intensive training model for both PMEs and Biofeedback, with office visits three times a week for four weeks. In addition, all patients received an educational "vaginal palpation" each week. The biofeedback group obtained a statistically significant improvement after only two weeks (6 training sessions), but after a total of four weeks, both groups leveled off at the same 55% symptom reduction (based only on pad weights[8]).

For PMEs only, 55% is a typical result. But for biofeedback, 55% sets a new record poor result. It is not obvious why Berghmans' biofeedback group did so much poorer than any other biofeedback group, so a closer examination is in order.

A. Randomization Issues.

Berghmans et al claim (Table IIIb) that there were "no significant differences" between the pre-treatment BF and PFMT groups except on fluid intake, but the PFMT group actually had 58% more leaks (diary, 3.0 vs. 1.9). In fact, Berghmans' PFMT group's incontinence, at 21 leaks per week, would have been classified as "severe" by Burns' (1993) standard (32+/-12 to 40+/- 18), whereas their BF group at 13 leaks would have been classified as only "moderate" (12+/-4 to 13+/-4). Clearly the PFMT group had much more room for improvement than the BF group.

B. Presentation of Results

Although the objective and subjective outcomes were very similar for both groups, it bears noting that in terms of the popular "patient improvement rates" (which are always higher than symptom reduction rates), the PFMT group had 85% cured or improved, while the BF group had 95% improvement. This suggests a trend for the superiority of biofeedback, in addition to the acknowledged time advantage.

C. Use of inappropriate equipment

Vaginal EMG was detected for biofeedback using a Verimed brand vaginal probe. Unfortunately, the selected probe is actually sold as an electrical stimulation device, not as a biofeedback sensor. It has two circular electrode bands, instead of longitudinal EMG sensors. Binnie et al (1991)[9] found that circumfirential electrodes, which are commonly used in stimulation, were very ineffective in detecting EMG signals. Thus Berghman's "biofeedback" subjects got very inferior feedback information, which may have contributed to their poor results with biofeedback.

Berghmans et al claim that the Myaction 12 biofeedback device "scores very well on reproducibility and validity" (p. 41) but it seems likely that applies only to conventional surface electrodes. The accuracy of the device coupled to an electrical stimulation probe has never been verified, and based on Binnie, it seems very unlikely.

D. Problems with The Biofeedback Protocol

Berghmans et al adapted the "intensive therapy model" (three clinical treatments a week) that Kari Bo had originally devised for manual physiotherapy. There is no prior biofeedback research using this protocol.

Apparently the exercise group was coached in doing exercises (in each of four positions) at each of three 30-minute thrice-weekly sessions, while the biofeedback group was coached in biofeedback (in one position) for the same amount of time. Then both groups were sent home with "identical exercise" instructions. But was their preparation really "identical"?

The Monaco summary (Table II) states that the Biofeedback group received "PFMT as above" (like the Control Group), but that assertion cannot be justified on the basis of the actual text, which claims that both groups were seen for "25-35 minutes per visit". Presumably that would have been done to "control" or equalize the amount of therapist contact time with each subject group. But it raises a difficult methodological issue, since the Biofeedback group had been given office practice with the biofeedback device, but were then sent home to do the "identical" set of exercises which they had not practiced under supervision. There is a very big difference between exercising with an inserted Verimed electrode (one of the largest probes ever made) in just one position, and exercising in four positions with an empty vagina. This difference could easily account for the lack of progress in the group that had not been fully instructed in home exercise.

Monaco classifies this report as "PFMT plus BF", but this seems to be a misnomer; in fact, Berghmans used "PFMT or BF" office training combined with PFMT home exercise.

2. The Burns et al study (1993)

Burns et al found only a trend towards better results in the biofeedback group (61% symptom reduction) compared with the PME group (54%) but the difference was not statistically significant.

While 54% Sx Reduction is typical of PME only studies, 61% is among the poorest results ever achieved in a biofeedback study --- that is, until Berghmans' work. Some of the reasons for the Burns' poor outcome are clear, and discussed at length in a review at http://www.incontinet.com/burns.htm. Some of the major problems are these:

1. Burns' over enthusiastic use of exclusionary criteria for research subjects resulted in an extremely disadvantaged patient population. She selected only 135 subjects out of 1,042 recruits, and ended up with a pool of older women with extremely weak muscles. In fact, her biofeedback subjects averaged only 2.0 microvolts before the intervention, whereas her own pilot study subjects had averaged 6.0 microvolts.

2. Burns' biofeedback protocol was an extremely truncated version of common biofeedback practice. At each weekly session she allowed only ten short contractions and ten 10-second contractions, for a total of about 5.5 minutes of actual training per week. In contrast, clinical practice in most clinics would be at least 30 minutes actual practice per office session, or more than five times what Burns offered. In addition, she did not do a standard muscle evaluation at each weekly session, so she did not have the opportunity to inform her patients of their progress (or lack thereof). Such objective measures are commonly considered an important part of the motivational aspects of effective biofeedback training.

3. Burns never evaluated the "dosage levels" that would be effective for the particular training procedures that she utilized; she merely assumed that eight weeks should be "right". Since no evaluations were conducted during the eight weeks, we have no way of knowing if her subjects had just begun to catch on, or had already plateaued, when the project was terminated after eight sessions.

4. Finally, the therapist who administered the biofeedback training sessions was never actually trained in the technique. Burns had stated in her grant application that this individual would become certified by the Biofeedback Certification Institute of America ("BCIA") before beginning the work. That was never done.

3. Castleden et al (1984)

The Castleden study should NOT have been included in this list, since their patients (unlike Shepherd, whose research they were attempting to replicate!) did not actually use the Kingsdown perineometer for "biofeedback-guided practice". Their subjects all did empty-vagina practice and only "used" the perineometer once a day (and then only for the first two weeks!) to "check" their contractions. It isn't clear what Castleden expected subjects to learn from this very unusual "check" procedure, which has never been reported before or since.

"Biofeedback" involves "real-time" feedback of muscle activity, according to experts in the field[10]. "Progress Reports" coming as much as 12 hours after the morning exercise session would not be expected to be very helpful, and this research actually proves that they are not.

In fact, Castleden's own report shows their intervention was not effective; their subjects' contractions only improved 2 cmH2O, compared with 13 cmH2O for Shepherd, whose work they were supposedly replicating. It is clear that this intervention ("daily checking") was NOT helpful. Since they didn't do biofeedback, they should not be included in the biofeedback list.

4. Glavind et al (1996)

Glavind et al (1996) is the only biofeedback RCT that obtained a statistically significant difference -- even at 2.5 years follow-up, their biofeedback group had 87% patient improvement rate, compared with only 21% for the PME only group. That improvement is not only statistically, but also clinically, significant as well. It is an exemplary study.

5. Taylor & Henderson, 1986

The Taylor and Henderson study was excluded by Monaco solely because of "very small sample size (n=12)" [p. 593]. This exclusion seems both prejudicial and inappropriate. (It would have been more appropriate to include it but weigh it in proportion to the group size.) This project was advanced as a pilot to a much larger study, and showed a decisive advantage for biofeedback. Unfortunately, the larger study was never undertaken, due to the untimely death of Dr. Taylor. It should be noted that the authors did not conduct any statistical tests because the superiority of the biofeedback was obvious, and they recognized the size limitations of their pilot project.

So, we have three biofeedback studies that fit the pre-established form for validity; one has a statistically significant outcome in favor of biofeedback, the next a trend for biofeedback, and the third shows a time-in- therapy advantage for biofeedback. From these three they conclude that there is NO ADVANTAGE to biofeedback, compared with PME alone. Does it make sense? Only to someone who's mind was already made up.

As mentioned, Berghmans et al did a "systematic review" article published in BJU in 1998 that reached the same "no advantage" conclusion that was subsequently adopted by the Monaco ICS/WHO conference.

But just recently, Mark Weatherall of New Zealand published in BJU (1999) a meta-analysis of the same five studies discussed by Berghmans et al. He writes:

"...because only 1 of the 5 papers supported BF over PFE alone [at a statistically significant level], this was taken [by Berghmans] as strong evidence that biofeedback (BF) did not offer anything in addition to PFE alone. 3 of the 5 papers [i.e., Berghmans, Burns & Glavind] gave sufficient quantitative data in their reports to be subjected to the meta-analysis using odds ratios and as I have stated in my paper in the BJU the pooled analysis just about favors BF and PFE against PFE alone, the opposite conclusion of the systematic review." [Personal email, July 6, 1999, quoted with permission. (See the full abstract included in References, below)].

THE KEGEL MODEL

Despite recognition of Kegel's historic role (p. 586) it bears noting that none of the four studies cited by the Monaco report as examples of "biofeedback" followed the original Kegel model, which involved using biofeedback for daily at-home practice. All of them followed instead the Burgio - NIA model of "office-only" biofeedback, with "empty vagina" practice at home.

Kegel's model called for 90+ biofeedback sessions a month. Berghmans used 12 biofeedback sessions in a month, Burns used 8 weekly sessions, and Glavind used only 4 weekly sessions. Burgio used an average of 4 weekly or biweekly office sessions.

6. Shepherd, Montgomery and Anderson (1983)

The only study that DID follow the Kegel "home practice with biofeedback" model was the one misclassified as a "resistive device" study in Monaco; Shepherd, Montgomery and Anderson (1983). These authors used a manometric perineometer similar to Kegel's device, and achieved a 300% increase in pelvic muscle strength. As a result, they obtained an 83% average group symptom reduction for BF+ PME, vs. only 25% for the control (PME only) group.

It is not really clear how the Committee 14 could make such a blatant mistake. The study is listed in the second part of table 2, "PFMT with intravaginal resistance devices (IVRD)...", but in the "intervention" column it is clearly stated "PFMT + IVRD (intravaginal pressure device with visual pressure biofeedback used daily at home." (Emphasis added)

INTRA-VAGINAL RESISTIVE DEVICES

The discussion of IVRD in the text is equally erroneous. The text states:

"PFMT with IVRD vs. PFMT alone. Both trials [76, 77 {i.e., Shepherd et al and Ferguson et al}] investigated the use of PFMT with IVRD versus PFMT ... and found significant improvement in both groups and no differences between the groups."

That conclusion is partly false and partly misleading. First of all, Shepherd et al did not find "no differences", but they found a 83% vs. 25% advantage for the biofeedback group. That is a substantial difference that has not only obvious statistical significance, but clinical significance as well.

Second, while Ferguson et al found a 61 vs. 47 percent advantage for the PME only group, there was a serious mismatch in their "randomization", which should have disqualified the study from consideration.

The resistive device group started out with 23.2 cmH2O contractions before therapy, compared with 38.3 for the control group. After treatment, the resistive device group improved 44% (to 33.4), while the control group only improved 21% (to 46.5). In other words, Ferguson's "control" group started out with stronger muscles than the experimental group ever obtained, even after therapy [p. 673]. Based on this mismatch, no conclusions can be properly drawn from this "resistive device" study.

Just as Weatherall's review of Berghmans came to an opposite conclusion, we conclude that a corrected version of the ICS/WHO Monaco section on biofeedback [still allowing their prejudicial exclusionary criteria] yields this:

Glavind, 1996	(n=40)	Stat. sig. advantage for BFB
Shepherd, 1983	(n= 22)	An obvious advantage for BFB
Burns, 1993	(n=120)	Trend favoring BFB
Berghmans, 1996	(n=40)	Time advantage for BFB
Taylor, 1986	(n=12)	Clear advantage for BFB
Total	(n=234)	All studies positive for BFB

The analysis of biofeedback research, even when artificially restricted to RCT studies, shows a clear advantage, in 100% of the studies, in favor of supplementing PME with BFB. The "no value" conclusion of the ICS/WHO report is not supported by the evidence.

If we accord to biofeedback the same considerations that Monaco accorded to other Behavioral Therapies, we must also consider the following well-respected Clinical Series research reports[11]:

Authors.	Population	Treatment Effects	Symptom Reduction	Patients Improved
Burgio, Robinson & Engel, 1986	22 Stress Incontinent women	N/A[12]	76% for biofeedback 51% controls	@>60%, 92% biofeedback 50% controls
Burgio, Stutzman & Engel, 1989	20 Post prostatectomy	N/A	81% Urge Inc. 78% Stress Inc.	N/A
Burgio, Whitehead & Engel, 1985	39 elderly outpatients	N/A	82% Stress, 85% Motor Urge, 94% Urge w/o d.i.	82% improved to <1 leak/week
Burgio, Locher, Goode, Hardin, McDowell, Dombrowski & Candib, 1998[13]	197 geriatric women in RCT	N/A	81% biofeedback[14], 68% drug, vs. 39% controls[15]	90% biofeedback, 77% drug, vs. 65% controls >50% improved
Burton, Pearce & Burgio, 1988	27 elderly Urge Inc.	N/A	79% biofeedback 82% behavioral	N/A
McDowell, Burgio & Dombrowski, 1992	29 elderly with multiple medical problems	N/A	82% biofeedback plus behavioral	34% cured N/A improved
Middaugh, Whitehead & Burgio et al, 1989	4 male stroke patients	N/A	100% biofeedback	100% cured

O'Donnell & Doyle, 1991	20 male inpatients >65 y.o. vs. 28 no-treatment controls	N/A[16]	61% biofeedback, -2% controls	N/A
Susset, Galea & Read, 1990	15 Stress Incontinent women	283% increase	86% reduction (pads)	80% cured + 13% improved >65% =93%
Baigis-Smith, Smith, Rose, Newman, 1989	54 women >60	100% increase	78% (end of treatment)	37% cured, plus 39% > 60% = 76%

OTHER ISSUES

Biofeedback experts have long complained that the AHCPR Guideline on Urinary Incontinence inappropriately misclassifies electrical stimulation as a sub-category of pelvic muscle rehabilitation (p. 31). The Monaco Report, in contrast, considers electrical stimulation to be distinctly separate from pelvic muscle rehabilitation and discusses it in a separate section and table.[17]

The HCFA TAC minutes of May 6-7, 1997 referred to verbal instruction in "Kegel exercises as "the time honored method" of pelvic muscle rehabilitation, a slogan that was echoed in the Continence Coalition's Utilization Parameter for Biofeedback. The Monaco report reminds us that pelvic muscle training with "a perineometer for resistance and biofeedback" was introduced in 1948 by Arnold Kegel, MD. Renewed interest in pelvic floor exercise for "urinary incontinence was not evident until the 1980's" (p. 586).

Thus biofeedback, and not verbal instruction, is the "time honored method" of treating incontinence.

References (Abstracts)

Baigis-Smith, J., D. A. Smith, et al. (1989). “Managing urinary incontinence in community-residing elderly persons.” Gerontologist 29(2): 229-33.

Abstract: This two-year project demonstrated a significant decrease over time in urinary accidents after instruction in Kegel exercises augmented by the use of biofeedback, habit training, and relaxation techniques in 54 cognitively intact volunteers aged 60 years and over who had stress, urge or complex types of incontinence. This decrease in urinary accidents per week was maintained from the end of focused treatment through 6-month and 1-year follow-up, despite the age of the participants, previous urinary-related surgeries, or duration of incontinence.

Burgio, K.L., W. E. Whitehead et al (1985). "Urinary incontinence in the elderly. Bladder-sphincter biofeedback and toileting skills training." Ann Intern Med 103(4): 507-15

Abstract: Behavioral treatment of urinary incontinence was given to 39 elderly outpatients; 19 had stress incontinence, 12 detrusor motor instability, and 8 urge incontinence without instability. Biofeedback involving the bladder and sphincter was used to teach selective control of sphincter muscles or voluntary inhibition of detrusor contractions. Traditional behavioral methods used included habit training to gradually increase the voiding interval and relaxation training to cope with the urge to void. After an average 3.5 training sessions, patients with stress incontinence reduced the frequency of incontinent episodes an average of 82% (range, 55% to 100%). Patients with detrusor motor instability showed an average 85% improvement (range, 39% to 100%), and patients with urge incontinence reduced incontinence an average of 94% (range, 83% to 100%). Furthermore, 13 of the patients achieved total continence, and 19 had fewer than one accident per week after treatment.

Burgio, K.L. , J. L. Locher et al.(1998) "Behavioral vs Drug Treatment for Urge Urinary Incontinence in Older Women." JAMA. 280:1995-2000

Context.—Urinary incontinence is a common condition caused by many factors with several treatment options. Objective.—To compare the effectiveness of biofeedback-assisted behavioral treatment with drug treatment and a placebo control condition for the treatment of urge and mixed urinary incontinence in older community-dwelling women. Design.—Randomized placebo-controlled trial conducted from 1989 to 1995. Setting.—University-based outpatient geriatric medicine clinic. Patients.—A volunteer sample of 197 women aged 55 to 92 years with urge urinary incontinence or mixed incontinence with urge as the predominant pattern. Subjects had to have urodynamic evidence of bladder dysfunction, be ambulatory, and not have dementia. Intervention.—Subjects were randomized to 4 sessions (8 weeks) of biofeedback-assisted behavioral treatment, drug treatment (with oxybutynin chloride, possible range of doses, 2.5 mg daily to 5.0 mg 3 times daily), or a placebo control condition. Main Outcome Measures.—Reduction in the frequency of incontinent episodes as determined by bladder diaries, and patients' perceptions of improvement and their comfort and satisfaction with treatment. Results.—For all 3 treatment groups, reduction of incontinence was most pronounced early in treatment and progressed more gradually thereafter. Behavioral treatment, which yielded a mean 80.7% reduction of incontinence episodes, was significantly more effective than drug treatment (mean 68.5% reduction; P =.04) and both were more effective than the placebo control condition (mean 39.4% reduction; P<.001 and P=.009, respectively). Patient-perceived improvement was greatest for behavioral treatment (74.1% "much better" vs 50.9% and 26.9% for drug treatment and placebo, respectively). Only 14.0% of patients receiving behavioral treatment wanted to change to another treatment vs 75.5% in each of the other groups. Conclusion.—Behavioral treatment is a safe and effective conservative intervention that should be made more readily available to patients as a first-line treatment for urge and mixed incontinence.

Burton, J. R., K. L. Pearce, et al. (1988). “Behavioral training for urinary incontinence in elderly ambulatory patients.” J Am Geriatr Soc 36(8): 693-8.

Research questions addressed by this study were: 1) Is the treatment of chronic urinary incontinence (UI) in elderly, nondemented ambulatory patients using bladder-sphincter biofeedback as effective when performed by an internist/geriatrician and a nurse practitioner as that reported by behavioral scientists?; and 2) how does bladder-sphincter biofeedback compare to a program of behavioral training that does not utilize biofeedback? Twenty-seven patients with UI were assigned based on the number of baseline accidents documented in a self-maintained log, their sex, and the predominant pattern of symptoms (urge or stress) to one of two treatment groups: biofeedback (13 patients) or behavioral training not utilizing biofeedback (14 patients). Patients were given up to six treatments. Patients in both groups achieved a highly significant (P less than .001) reduction in urinary accidents 1 month following treatments compared with their baseline number of accidents. The average reduction of accidents over this time period was 79% for the biofeedback group and 82% for the group receiving behavioral training without biofeedback. All patients showed improvement and no patient experienced any side effect. A internist/geriatrician and a geriatric nurse practitioner may achieve success utilizing behavioral therapy with or without biofeedback for the treatment of chronic urinary incontinence for ambulatory elderly patients.

McDowell, B. J., K. L. Burgio, et al. (1992). “An interdisciplinary approach to the assessment and behavioral treatment of urinary incontinence in geriatric outpatients.” J Am Geriatr Soc 40(4): 370-4.

OBJECTIVE: To test the effectiveness of an interdisciplinary assessment and behavioral treatment of persistent urinary incontinence in geriatric outpatients. DESIGN: Prospective case series in which frequency of incontinence was measured before and after intervention. SETTING: We established an interdisciplinary continence program within an existing academic center, the Benedum Geriatric Center. PATIENTS: Convenience sample of 70 non-demented outpatients aged 56 to 90 years. Behavioral treatment was provided to 29 patients including many with multiple medical problems (Mean = 6.0 problems). INTERVENTION: Behavioral treatment consisted of biofeedback, pelvic floor muscle exercise, scheduled voiding, and other strategies for preventing accidental urine loss. OUTCOME MEASURE: Outcome of treatment was measured by comparing bladder diaries completed in the 2 weeks immediately following treatment to those completed in the pretreatment phase. RESULTS: Following an average 5.6 treatment sessions, the mean weekly frequency of accidents was reduced from 16.9 to 2.5 (P less than 0.01). Individual reductions ranged from 30.8% to 100% with an average of 81.6% improvement. Ten patients achieved continence. Patients with mixed incontinence had greater improvement than those with urge incontinence alone (P less than 0.05), and patients who reported previous evaluation or treatment had a poorer outcome than those coming for their first evaluation (P = 0.05). Degree of improvement was not significantly related to age, duration of symptoms, baseline frequency of accidents, number of treatment sessions, number of other medical diagnoses, or urodynamic findings. CONCLUSION: We conclude that older adults who are able and willing to participate in behavioral treatment can benefit significantly despite other health problems or disabilities.

Middaugh, S. J., W. E. Whitehead, et al. (1989). “Biofeedback in treatment of urinary incontinence in stroke patients.” Biofeedback Self Regul 14(1): 3-19.

Abstract: Urinary incontinence can occur poststroke owing to weakness or incoordination of sphincter muscles, impaired bladder sensation, or hyperreflexic, neurogenic bladder. Four male subjects who had urinary incontinence associated with a stroke that had occurred 8 months to 10 years earlier, and who averaged 1.6 to 7.5 accidental voidings per week, participated in an outpatient study with a 4-week scheduled- voiding baseline, 2 to 5 sessions of biofeedback-assisted bladder retraining, and 6- to 12-month follow-up. Training sessions included stepwise filling of the bladder and manometric feedback display of bladder pressure, abdominal pressure, and external anal sphincter pressure. Training procedures were designed to teach subjects to attend to bladder sensations, inhibit bladder contractions, and improve voluntary sphincter muscle control. All four subjects achieved and maintained continence regardless of substantial differences in subject characteristics, including laterality of stroke, degree of sensory impairment, and independence in daily activities.

O'Donnell, P. D. and R. Doyle (1991). “Biofeedback therapy technique for treatment of urinary incontinence.” Urology 37(5): 432-6.

Biofeedback treatment of urinary incontinence is a management method that has low risk and therapeutic efficacy for selected patients. Biofeedback therapy techniques vary widely and have not been well described or standardized. A technique for biofeedback therapy is described that allows accurate signal monitoring and assures appropriate biofeedback to the patient. External anal sphincter electromyographic performance is presented to the patient as a color line graph with pitch variable audio feedback. The method has complete flexibility in providing biofeedback training according to patient performance level and is one that can be easily interpreted by patients who have voiding dysfunctions.

Susset, J. G., G. Galea, et al. (1990). “Biofeedback therapy for female incontinence due to low urethral resistance.” J Urol 143(6): 1205-8.

Abstract: Urinary incontinence, mostly secondary to low urethral resistance, in 15 women was treated for 6 weeks by biofeedback. A new device equipped with visual and audio signals connected to an intravaginal probe was used by the patient for 15 minutes twice a day. Of the patients 12 were continent subjectively and objectively, 2 had 65 and 75% improvement and could lead a normal life, and only 1 failed to respond and was treated surgically. Besides the quality of the device, success depends largely on the quality of moral support given to the patient during the treatment.

Weatherall M. (1999) "Biofeedback or pelvic floor muscle exercises for female genuine stress incontinence: a meta-analysis of trials identified in a systematic review." BJU Int. Jun; 93(9): 1015-6.

ABSTRACT: OBJECTIVE: To test, by meta-analysis, the conclusion of a systematic review that biofeedback was no more effective than pelvic floor muscle exercises alone for the treatment of female genuine stress urinary incontinence. MATERIALS AND METHODS: Data extracted from the five trials identified in the systematic review were subjected to pooled analysis of odds ratios for the outcome of "cure". RESULTS: The odds ratio for biofeedback combined with pelvic floor muscle exercises, compared with pelvic floor muscle exercises alone, leading to cure was 2.1 (95% confidence interval 0.99-4.4). CONCLUSIONS: Biofeedback may be an important adjunct to pelvic floor muscle exercises alone in the treatment of female genuine stress urinary incontinence. A qualitative statistical analysis of the studies identified leads to a different conclusion from those of the systematic review.

---

[1] Report of the Ad Hoc Committee to Suggest Broad Guidelines for the Management of Urinary Incontinence, Alan J. Wein, MD, chairperson. (Baltimore: AUA, 1999).

[2] Sponsored by the World Health Organization (WHO) and the International Union Against Cancer (UICC) Published (1999) as Incontinence, by Paul Abrams, Saad Khoury, and Alan Wein (eds), ISBN 1 898452 25 3, Plymbridge Distributors, Ltd, Health Publications Lts, UK.

[3] From the report of Committee 14, Conservative Management in Women; P.D. Wilson, Chair; Members K. Bo, A. Bourchier, J. Hay-Smith, D. Staskin, I. Nygaard, J. Wyman; consultant, A. Shepherd

[4] In the USA "RCT" refers to "randomized controlled trials"; see, e.g., the AHCPR Guideline (1996), page 40. Research omitted by Monaco (e.g. Burgio, Robinson & Engel, 1986) is described by the Guideline as "a controlled study". Apparently the European usage is slightly different.

[5] The only exception was the inclusion, in the text only, of a single "controlled clinical trial" by the chairman of Committee 14, P. Wilson.

[6] Addition, a very different MEDLINE search was used; for Physical Therapies, 1980 to November 1997 was used; for Bladder Retraining, 1966 to February 1998 was reviewed, or 17 years vs. 32 years.

[7] The five RCTs were later expanded in Table 5 to include a total of ten (10) RCTs. Table 5 (in a serif typeface) was apparently prepared separately from the others (which are all in a sans-serif typeface).

[8] Apparently Incotest pads were used for 48 hours at the start, middle and end of therapy only.

[9] Binnie et al, "The importance of the orientation of the electrode plates in recording the external anal sphincter EMG by non-invasive anal plug electrodes. Int. J. Colorectal Dis, 1991, 6:5-8. Circular electrodes detected only 38 to 46% of the EMG signal detected by longitudinal electrodes (p. 6). [Binnie.htm]

[10] See, for example, Schwartz, M. Biofeedback. Guilford Press, 1995

[11] We have added a column called "Treatment Effects", which are commonly reported in biofeedback research to demonstrate whether or not the intervention had the desired physiological effect, in this case, increasing the "strength" of a pelvic muscle contraction. "Symptom Reduction" and "Patients Improved" are measures of "Outcome Effects", i.e., addressing the question of whether the intervention changed the offending symptoms.

[12] Manometric biofeedback systems such as those used by Burgio and the NIA produce relative-scale data that is orifice-size dependant and has no absolute value; it is, therefore, seldom reported. It is, however, traditionally reported in fecal incontinence research, despite the meaninglessness of the "data" on any absolute scale.

[13] Note that Burgio's JAMA article was published six months after the Monaco meeting.

[14] Biofeedback was limited to one (74% of subjects) or two office training sessions only.

[15] From the Comments section: "The control condition should not be interpreted as a no-treatment condition, since like the other groups, they were active participants in treatment. Control patients consumed capsules that they knew could have contained the medication, completed detailed bladder diaries throughout the 8-week intervention phase, attended 4 clinic visits, completed an adverse effects checklist at each visit, and received therapeutic attention from a nurse practitioner who reviewed the diaries with them and inquired about their progress and concomitant events." Historically, "behavior monitoring" (detailed recording of one's own behavior), is well known in psychology to be therapeutic in itself, so it is not surprising that it produced a substantial effect (almost 50% of biofeedback's effects).

[16] O'Donnell used the EMG channel of a urodynamics system, and did not report objective measures of physiological variables.

[17] Hofbauer et al, 1990, is not an exception to this, since Hofbauer actually did teach pelvic floor muscle training in two experimental groups in addition to using electrical stimulation in three groups (p. 620).